U.S. patent number 7,310,143 [Application Number 10/981,801] was granted by the patent office on 2007-12-18 for nist traceable automated visual inspection system for an inspection of particles in solution.
Invention is credited to Gerald Walter Budd.
United States Patent |
7,310,143 |
Budd |
December 18, 2007 |
NIST traceable automated visual inspection system for an inspection
of particles in solution
Abstract
A method for the substantially complete detection and
measurement of all particles, within a predetermined size, range,
contained in an injectable solution comprising the steps of: a)
rotation of the container causes substantially all of the particles
in the injectable solution in the container to be set in motion; b)
uniformly illuminating the background around the container with
light; and c) detecting at least one of light scatter, light
reflection and light extinction caused by said particles, with
detectors having a depth of focus of detection in a specified
volume of the container. Wherein the detectors are positioned,
relative to the container whereby the optical path and field of
view allows the sensor sufficient focus to view substantially all
of the bottom interior surface of the container and substantially
all of the solution volume within the container. The method and
apparatus produces a geometric representation of the particles in
the detection region, whereby the size of detected particles can be
is accurately adjusted to an actual size by either calculation or
by calculated offset to allow accurate measurement of particle
dimensions.
Inventors: |
Budd; Gerald Walter
(Farmington, MI) |
Family
ID: |
34556467 |
Appl.
No.: |
10/981,801 |
Filed: |
November 5, 2004 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20050099625 A1 |
May 12, 2005 |
|
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
60518699 |
Nov 9, 2003 |
|
|
|
|
Current U.S.
Class: |
356/335;
356/237.1; 356/240.1; 356/427 |
Current CPC
Class: |
G01N
15/0205 (20130101) |
Current International
Class: |
G01N
15/02 (20060101) |
Field of
Search: |
;356/335,427,428,237.1,238.1,240.1 ;250/223B,208.1 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Toatley, Jr.; Gregory J.
Assistant Examiner: Nur; Abdullahi
Parent Case Text
CROSS-REFERENCE TO RELATED APPLICATIONS
I claim priority to my Provisional Application No. 60/518,699 with
filing date Nov. 9, 2003.
Claims
What is claimed is:
1. An improved apparatus for the detection of contaminating
particles in the fluid of small containers in which the apparatus
produces a uniform illumination field for consistent grayscale
detection and measurement of moving particles in the solution using
a machine vision measuring system comprising: a) an image
processing computer for image acquisition, image storage and image
processing capability; b) the image processing computer comprising
memory for storing the images formed by the camera; c) the image
processing computer also comprising digital parallel input/output
digital serial, and Ethernet communication capabilities for
providing messages to external devices to report one or more
measurements or characteristics of the particles moving in the
solution; d) the image processing computer executing control
software stored in a computer readable medium, for allowing request
and response signals from external devices indicating a small
container to be inspected, for causing the image processing
computer to perform image analysis for extraction of the summation
of grayscale values of the moving particles in the solution found
within the small container, as well as for causing the image
processing computer to store a reference images of an acceptable
quality of small container with no presence of contamination
particles in a memory location referenced by a specific
identification code that is unique to a small container size and
shape with a specific fluid fill level; e) an image sensor with
appropriate lens for providing a spatial resolution and depth of
field necessary to form a sharp focus image of substantially all of
the bottom interior surface of the small container; f) wherein the
image sensor comprises sufficient pixel resolution to resolve a
contaminating particle of at least 40 micrometers diameter resting
on the interior bottom center of the small container; g) wherein
the image sensor will acquires images at a rate of 20 to 30 frames
per second; h) the image sensor and optical components are mounted
inside a sealed enclosure with an optical window so that the image
sensor can view objects outside the enclosure without obstruction;
i) the optical components are aligned so that the optical path is
at a downward angle less than perpendicular from the axis of
rotation permitting the sensor to view substantially all of the
bottom interior surface of the small container; j) an illumination
system comprised of a cube structure with a "U" shaped channel cut
into a cube of optical grade polycarbonate or acrylic so that the
center of radius of the curved portion of the "U" shaped channel is
aligned with the axis of rotation of the container; k) an
illumination system with the "U" shaped channel providing uniform
diffuse illuminated surface large enough to allow the diameter of a
small container to fit inside without interference with the walls;
l) the illumination system implements multiple light emitting
diodes (LED's) arranged around three side of the diffusing cube and
positioned to uniformly illuminate an cylindrical object place at
the center of the "U" shape channel; m) the illumination system
utilizing one or more power supplies with control circuitry to turn
on or off sections the LED's as required by the image processing
system to enhance the contrast of the particles in the solution of
small containers; n) a precision drive motor is connected directly
to the recessed bottom holder is used to impart rotational motion
to the base of the small container with a motion processing
computer and motor drive unit comprising memory for storing one or
more defined motion programs; o) the rotational motion is limited
to insure that angular acceleration and velocity do not deform the
meniscus to cause cavitation of the fluid or creep up the walls to
the neck region of the container; p) the rotational motion of the
container imparts a motion to all particles within the fluid
contents of the container; q) the image sensor will begin to
acquire images at a predefined acquisition rate after the
rotational motion of the container has stopped, as images are
acquired they are stored the image processing memory for analysis
after the acquisition of all required images are completed, a
minimum of four images are required for analysis but typically are
20 or more are used; r) whereby the image sensor is positioned
relative to the axis of rotation of the small container, whereby
the focal point of detection coincides with the axis of rotation of
the small container so to view the content of the solution in the
small container and specifically the interior bottom of the small
container with the illumination system provide a contrasting
geometric shape of substantially all of the contaminating particles
in the solution being identified and reporting grayscale
information of each particle being displayed on a human machine
interface.
2. The apparatus as claimed in claim 1, wherein the illumination
cube has a clam shell design with an upper and lower shell that may
be disassembled to allow access to the interior components for easy
assembly and repair.
3. The apparatus as claimed in claim 1, wherein the illumination
source of the illumination cube is comprised of flat panel LED's
constructed with the LED's in close proximity to each to provide an
extremely uniform illumination field.
4. The apparatus as claimed in claim 1, wherein the LED's utilized
in the illumination cube may be energized in whole or in sections
to illuminate the contents of the container being tested from
various directions to provide backlighting, side lighting or a
combination or both.
5. The apparatus as claimed in claim 1, wherein the diffusing
element of the illumination cube is constructed of single block of
optical grade polycarbonate or acrylic that is machined with an
elongated "U" shaped cutout parallel to the upper and lower
surfaces so that the major axis of the container is aligned with
the radius of curvature of the "U" shaped cutout and thus providing
illumination on all sides of the container with the exception of
the viewing direction.
6. The apparatus as claimed in claim 1, wherein the sensor module
is constructed with a folder optical path design through the use of
mirror to extend the distance between the sensor and the container
under observation to improve the depth of field while minimizing
the physical separation between the sensor and the rotational axis
of the container.
7. The apparatus as claimed in claim 1, wherein the sensor module
may utilize one or more optical filters in the optical path between
the sensor and container positioned for inspection to enhance the
contrast of the contaminating particles in the solution.
8. The apparatus as claimed in claim 1, wherein the optical path of
the sensor module is oriented at a downward angle relative to the
axis of rotation of the container to allow substantially all of the
bottom interior surface of the container to viewed by the sensor
permitting the image capture of heavy contaminating particles that
may be lying on the bottom surface on the container.
9. The apparatus as claimed in claim 1, wherein the optical path
between the sensor and the container to be inspected and the
viewing angle allow a large percentage of the fluid contents on the
container to be viewed with each image acquisition and
substantially the entire contents to viewed in four images.
10. The apparatus as claimed in claim 1, wherein the viewing angle
of the sensor is oriented at a downward with respect to the axis of
rotation with sufficient field of view allow the fluid contents of
the container to be inspected from the bottom of the meniscus to
the bottom interior surface.
11. The apparatus as claimed in claim 1, wherein the high density
spacing of the LED's used in the illumination cube provide uniform
energy of illumination so a contaminating particle produces
approximately equivalent grayscale summation at all locations in
the fluid volume with exception of the extreme edges of the
container.
12. The apparatus as claimed in claim 1, wherein the precise
control of the rotational motion of the container by the drive
system is gentle and imparts motion to contaminating particles in
the solution without causing violent distortion of the meniscus of
cavitation within the fluid the container.
13. The apparatus as claimed in claim 1, wherein the precise
rotational motion control and the ability to acquire images allows
the user to study the shape of the meniscus and precisely define
the parameters for a controlled particle agitation for a specific
container size, shape and fluid fill level to insure reliable
detection of contaminating particles.
14. The apparatus as claimed in 1, wherein the combination of
illumination, sensor and precise motion control for excitation of
the particle movement within the fluid of the container under
inspection allows the generation of consistent grayscale images
that may be compared as part of a sequence of images to identify
differences between images as contaminating particles. The
summation of the subtle grayscale differences can be recorded and
compared with standard samples prepared with known NIST traceable
particles of various sizes. A plot of grayscale summation
difference versus NIST particle size will yield a calibration curve
that can be used to estimate the size of an unknown particle. A
separate calibration curve can be constructed for containers
different shapes, sizes and fluid fill levels.
15. The apparatus as claimed in claim 14, wherein a calibration
curve can be used for the rejection of unacceptable product based
on size of the contaminating particle rather than the simple
detection of a contaminating particle.
Description
FIELD OF THE INVENTION
This invention relates to the procedures and devices utilized in
the optical inspection of transparent containers for the presence
of contaminating particulate matter and particularly to inspection
of injectable pharmaceutical preparations.
BACKGROUND OF THE INVENTION
There is a legal obligation by manufacturers of pharmaceutical
injectable solutions to ensure that the product is free of
`visible` particle contaminants prior to their clinical use. This
legal obligation can be satisfied by the use of a labor intensive
and costly 100% manual inspection of injectable solutions. Less
costly automated particle detection systems have been developed.
However, in order to satisfy Good Manufacturing Practice, automated
inspection systems must be validated prior to any pharmaceutical
use. In the validation demonstration, the functioning of the
automated system must be shown to be at least as effective in
detecting and rejecting containers with `visible` contaminating
particles as the preceding manual inspection.
The performance of human `visible` particle inspection has been
characterized in published reports as a probabilistic process
without a sharp particle size accept/reject decision threshold
(i.e., a soft decisional process). In the production of an
injectable product under good control, the distribution of
contaminating particles is approximately hyperbolic, with the
concentration of contaminating particles decreasing rapidly as
particle size increases. The effect of the `soft` accept/reject
decision threshold is that a proportion of particle-contaminated
containers that should be rejected are accepted. A false reject
rate of good containers also results from the `soft` accept/reject
decision process. Due to the increased number of containers with
particles well below both clinical and control interest, a
disproportionate number of the containers that should be accepted
are rejected. This disproportionate false reject rate imposes
additional costs on the quality assurance program.
Validation of alternative equipment or methods is a Good
Manufacturing Practice requirement. The validation of a
contaminating particle inspection system is a demonstration that
the automated inspection system rejects those containers identified
in a manual inspection to be contaminated with "visible" particles.
It must show that the rejection capability of the automated system
is at least equal to or better than that achieved by the preceding
human inspection method. This demonstration must be successfully
completed prior to any production use of any proposed automated
system.
This demonstration is based on an established statistically
evaluated human `visibility` performance benchmark. To make
possible statistical comparisons and evaluations of particle
contamination, an inspection model was defined with a statistically
described rejection zone boundary. As currently accepted in the
pharmaceutical field the Reject Zone includes the group of particle
contaminated containers rejected in 70% of a series of manual
container inspections. The group of containers with a manual
rejection probability equal to or greater than 70% constitute the
"must reject" visible particle contaminated group.
Holographic measurements found that the size of the contaminating
particles that resulted in the 70% reject rate was 100. mu.m. This
determination was made with the particle contaminated containers
that were rejected in a 17 second, timed single container
inspection performed under 225 foot-candles of illumination, the
inspection time is equally divided against a black and white
background. The holographic data was correlated with the
statistically evaluated probability of detection data to define the
minimum `visible` particle size of 100. mu.m. Accordingly in
present practice all containers with 100. mu.m or larger
contaminating particles: are considered to be `must rejects`.
This Reject Zone definition has become a de-facto world standard in
validation demonstrations and any proposed automated inspection
device must function with at least the capability of the preceding
manual inspection. This equivalent functionality is demonstrated by
the achievement of an equal or higher rejection rate for the
containers identified in the manual inspection to have `must
reject` contaminating particles that are 100 .mu.m or greater.
When current commercially available automated inspection systems
were evaluated according to this standard, it was determined that
none could demonstrate, in a single inspection, results as secure
or as selective as that achieved by human beings. The proportion of
"must-reject" containers rejected in a single automated inspection
is between half and two thirds that of a skilled human
inspector.
As a result, in order to validate these automated inspection
systems (to match their inspection security to that of the
preceding manual inspection), a two inspection sequence is
currently employed. Only containers accepted in both inspections
are accepted for stock. Containers rejected in either of the two
sequential inspections are eliminated.
It has been determined that the limiting particle
rejection/detection probability for an inspection system is the
proportion of the liquid contents that have been examined for
particulate contamination. A complicating factor is that the
position of a contaminating particle in a container at the start of
each inspection is completely random. This random initial particle
position results in random distribution of particle orbits and
velocities within the container. The random particle velocity
distribution ranges from zero-to some design maximum.
A defined velocity of particle movement is employed to distinguish
between contaminating particles and stationary container markings
and optical defects. Particles that do not traverse the fractional
inspected volume or that move with insufficient velocity are not
detected. To improve the inspection security results, the
two-inspection `game of chance` technique to reduce the effect of
the random particle position and velocity is employed. Application
of classical probability theory shows that particle detection
security is enhanced but the discrimination of the accept/reject
decision compared to manual inspection is impaired when this
inspection technique is employed. The cost for this improvement in
detection probability is a four to six fold increase in the false
rejection rate of the manual inspection.
Ideally, secure detection, sizing and identification of the
contaminating particulates are an essential part of the control of
the production of pharmaceutical injectable products. However,
secure detection of randomly occurring and randomly positioned
particles in sealed transparent containers requires inspection of
the full volume of the container. In addition, accurate particle
sizing in the present automated inspection systems requires sharp
particle images. However, with present art, the sharp image
requirement cannot be achieved for the size range of containers
used for pharmaceutical injectable products.
In addition, only a portion of the contents of the container volume
is normally inspected for contaminating particles and accordingly
the security with which `must reject` containers are rejected in
the partial container volume inspection cannot exceed the
proportion of the container volume containing contaminating
particles inspected.
U.S. Pat. No. 3,627,423, issued Dec. 14, 1971, discloses an
improvement in particle contrast, and thus detectability, that
results from the use of narrow aperture lighting of the liquid
volume contents of the container. This patent teaches that narrow
aperture lighting of the liquid volume contents of the container
that transits the glass envelope or the container in a near
perpendicular condition minimizes the reduction in particle
contrast that occurs when a broad area light source is employed for
the inspection. The use of narrow aperture lighting of the liquid
volume contents of the container to produce forward scatter
lighting also minimizes the reduction of particle signal dynamic
range that occurs when glare reflections occur at the meniscus or
the container bottom. Glare reflections are produced when a bottom
mounted light source parallel to or on the container axis is
employed for the inspection. The teachings of this patent indicate
that measurements near the meniscus or the bottom of the container
are less sensitive.
At present there are two automated inspection methods, U.S. Pat.
No. 5,365,343 ('343 patent) issued Nov. 15, 1994, and U.S. Pat. No.
6,498,645 ('645 patent) issued Dec. 24, 2002, by the present
inventor that can equal or surpass the two important attributes of
the human inspection for contaminating particles in sealed
containers (the teachings of this patent are also incorporated
herein by reference thereto). These attributes are the reliability
of detection of these contaminating `visible` particles and the
selectivity of the human accept/reject inspection characteristic.
Both attributes are evaluated with statistical measures derived
from the probabilistic analysis of human inspection results.
In the '343 patent, an imaging lens is used at its maximum energy
collecting capability and its maximum resolution to achieve maximum
particle detection depth. Two light sources are employed, a forward
scatter light source is used for small and low contrast particle
detection. A second collimated light source, with intensity at the
detection, plane ranging from 0.2 to 10%, is used as a back
lighting means. The contaminating particles are sized numerically
by the peak change, either positive or negative, in light flux
collected from the moving particle. This patent teaches that the
light flux collected from an image and its blur surround is
essentially constant for a controlled displacement around the plane
of best focus.
In the '645 patent the measurement approach avoids reliance on
sharply defined image edges to detect and size particles, and it
results in a total light flux particle measurement. It relies,
however, on the presence of uniform illumination level for the
inspected container and system measurement stability. This reliance
results in particle detection variability determined by the
variation in the realizable illumination uniformity of the
inspected container and variation of the detection capability of
the system. The use of the light flux sizing as described in the
'645 patent makes it possible to inspect the full volume of a
container up to 30 mm in diameter with a 75 mm focal length lens at
maximum aperture of f stop equal to 1.8. The previous detection
volume limit was imposed by detection volumes 1 to 3 millimeters
thick centered on the axis of the container and extending through
its liquid contents. Since the reliability of detecting particles
in a container is proportional to the total container volume
inspected, inspection reliability for containers up to 30 mm in
diameter approaches 100% with the use of the teachings of the '645
patent. Determination of the size of a detected particle is
achieved with a stored transfer curve of particle size versus the
light flux peak detected. The methodology requires both light
source and measurement system stability to maintain the calibrated
particle sizing accuracy. Particles are detected by the variation
of light level received in each element of the photo detector. Any
change in the stability of the light source or the measurement
system affects the peak value of the detected light flux due to a
particle and thus the particle sizing accuracy. The approach
described in the '645 patent sacrifices particle image shapes to
achieve secure detection of the particle signal throughout the
volume of the container.
The use of the described present invention provides a uniform
illumination field within the volume of a container that permits
the detection of a contaminating particle in 90% of the solution
volume with a single image acquired by a photo detector (CCD
Camera). The present invention produces the uniform illumination
field using a uniquely shaped light emitting diode (LED) array
along with special diffusing element that surrounds the container
on at least 3 sides. The methodology produces a background
illumination that enhances the detection of contaminating particles
and allows the trajectory (position within container) of
contaminating particles to be mapped. With the capture of
successive images the invention provides nearly 100% detection of
contaminating particles contained in the solution. The image
processing technique uses a software algorithm to normalization
(reduce localized variations) in the image background. The result
of the uniform illumination on the image is to minimize variations
in the calculated size of the contaminating particles. In addition,
the invention provides geometrically correct images of
contaminating particles that may be accurately, sized when
positioned with specific inspection zones. The size of the
contaminating particles are determined by comparison of the pixel
dimensions of a particle to the pixel dimensions of previously
collected sample container(s) seeded with a single NIST traceable
particle. The present invention allows for the generation of
standard calibration curve for the determination of actual size of
contaminating particle (diameter in .mu.m) verses the apparent
particle size (diameter in pixels).
BRIEF SUMMARY OF THE INVENTION
It is an object of the present invention to transform the present
probabilistic detection of contaminating particles present in a
container, even larger than 30 mm in diameter, into a deterministic
detection and accurate measurement process
It is a further object of the present invention to provide a method
to generate a uniform illumination field within liquid volume in a
container to enhance the detection and measurement of contaminating
particle(s).
It is a further object of the present invention to provide a method
that evaluates the focused or nearly focused image of particle
measurement with a direct, physically based particle size
evaluation in a defined area.
It is a further object of the present invention to provide a method
to acquire image(s) of heavy contaminating particle(s) that are
positioned on the bottom of the container.
It is a still further object of the present invention to provide a
method that transforms the present random array of particles within
a container into a positioned array in a defined portion of the
container to be inspected with sufficient spatial resolution for
the accurate determination of size and/or shape.
It is yet another object of the present invention to provide a
means for the construction of an accurate instrument calibration
curve that will correlate actual size of NIST traceable particles
to the apparent dimension in the image sensor (pixels)
Generally the present invention comprises an improved method for
the substantially complete detection of all particles, within a
predetermined size range, contained in an injectable solution, in a
transparent container. In preferred embodiments the container has a
circular cross section, though some containers may depart from
circular symmetry in less preferred embodiments. The method
comprises the steps of:
a) pre-positioning particles in the container whereby rotation of
the container causes substantially all of the particles in the
injectable solution in the container to rotate, with approximately
equal initial velocity, in a shell volume adjacent the inner walls
of the container, with said shell volume having a predetermined
thickness;
b) illuminating all the particles rotating within the shell volume
with lighting means;
c) detection of particles by movement on the container bottom and
in solution by orienting the sensor with a downward angle with
respect to the axis of symmetry of the container;
d) detecting at least one of light scatter, light reflection and
light extinguishing caused by said particles, with detector means
having a depth of focus of detection in which said particles remain
in near-focus within the volume of the container; and
e) measuring at least one of light scatter, light reflection and
light extinguishing caused by said particles, with detector means
having a depth of focus of detection in which said particles remain
in focus within the center volume of the container.
wherein the sensed signal is corrected for the asymmetries of the
imaging system by correction means either by computation or by
repositioning the detector means relative to the container, whereby
a focused imaging plane is formed at the container axis and then
mechanically or electro-mechanically offset closer to the imaging
sensor than the center of the cross section, whereby the size of
detected particles in the opposite volumes is accurately
mathematically compensatible to an actual size. The lighting means
provides a multiplicity of directed light emitting diodes (LED's),
mounted on three of the interior walls of a cubic structure with an
acrylic element placed in the center. The diffusing element has a
"U" shaped channel removed along the centerline of cubic structure
in which the container being tested is positioned. The "U" shaped
diffusing element is designed to uniformly diffuse the light
entering the container. The sample container rest in slightly
recessed pocket centered on a rotational device. The center of the
axis of rotation is positioned to coincide with the center of
radius and width of the "U" shaped channel in the diffusing
element. The channel width of the diffusing element should be
approximately 1.5 times the diameter of the container being
inspected. The lighting means may be adjusted to enhance the image
characteristics by activating various LED lighting elements within
the structure. Contaminating materials with less optical density
can be enhanced in the image by reducing the radiant energy of the
illumination system.
With said detector being mounted inside a sealed enclosure the
critical optical components of the system can be protected from the
environment. The detector is mounted in such a manner so that so
that the optical path can be easily adjusted with the target area.
The design of the sensor enclosure allows for the insertion of
optical filter elements within the optical path of the
invention.
These and other objects, features and advantages of the present
invention will become more evident from the following discussion
and drawings in which:
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
FIG. 1 illustrates the exterior of the Illumination Module as
viewed from the side;
FIG. 2 illustrates the exterior and a portion of the interior of
the Illumination Module as viewed from the front or sensor viewing
direction;
FIG. 3 illustrates cross-section (A-A) of the Illumination Module
interior as viewed from above. The section cuts through the sample
container to show relative position of container with respect to
the illumination diffuser;
FIG. 4 illustrates cross-section (B-B) of the Illumination Module
interior as viewed from the side. The section cuts through the
sample container to show relative position of components;
FIG. 5 illustrates the front view of sensor enclosure;
FIG. 6 illustrates the interior of the sensor enclosure and
relative position of key components;
FIG. 7 illustrates the rear view of the sensor enclosure;
FIG. 8 illustrates the relative position of sensor enclosure with
respect to the sample container and spin access of rotational
device;
FIG. 9 illustrates the inter connections between the key components
of the invention;
FIG. 10 illustrates a multi station configuration of the
invention;
FIG. 11 is a graph showing the linearity of a plot of the apparent
size of the particle as determined by the invention plotted against
the actual size of NIST traceable single seeded samples.
DETAILED DESCRIPTION OF THE INVENTION
The invention is a combination of three key components configured
in the proper way to determine the maximum dimension of particles
in solution. The key components are an Illumination Module, a
Sensor Module, and an Image Processing System with specific
software. The invention has offers several unique components that
allow the particle to sized accurately. The invention has many uses
but is designed primarily for the detection of contamination in
clear solutions like those used in the pharmaceutical products.
The first key component is a unique illumination system designed to
provide a very uniform background for the inspection of product in
cylindrical vessels such as pharmaceutical vials. The illumination
system is cube shaped with a channel slightly larger than the
diameter of the vessel removed from the center, hereafter we shall
reference to this system as the illumination module. The basic
configuration is illustrated in FIG. 1. The cube is constructed
using an upper and lower halves indicated by items 1 & 2. The
construction is from a solid piece of aluminum that has material
removed to hollow its inner. The aluminum is anodized black to
insure that no reactive surfaces are on the components. The sample
product (pharmaceutical vial with liquid contents) is centered on a
recessed puck and held in position by a spring loaded clamping
device. Item 3 in the illustration represents the retaining sleeve
for spring and alignment shaft.
FIG. 2 illustrates the front view of the illumination cube with the
channel exposed. The sample product (item 14) is positioned on a
recessed bottom holder (item 7). The cap of the sample product
(item 5) is usually constructed of a rubber liner (cap) and a
protective aluminum closure. The clamping device used to securely
hold the sample container during rotation also has a recessed cup
in the contact area to center the sample (item 4). The clamping
device incorporates ball bearings to insure that the closure on the
sample is not damaged. The recessed bottom holder has two different
recessed diameters on the top and bottom surfaces. The recessed
holder is held tightly during rotation of the drive mechanism (item
8 and rotational drive of FIG. 1) using three equally spaced pins.
The inspection window (item 6) is centered in the most uniform area
of the illumination field. The illumination field is made uniform
by properly shaping the diffusing media and adjusting the LED
lighting sources.
FIG. 3 is the top view of the illumination module as seen through
section A-A of FIG. 1. The Aluminum housing (item 9) is hollowed
out to leave only a thin wall. Placed around the three walls
opposite the opening, are flat panels light emitting devices
(LED's). The LED flat panels are fabricated with a high density of
LED's per unit area, reference Phoenix Imaging 4100 series LED
backlights. The LED panels provide a uniform illumination and can
be turned on or off as required for the inspection. The uniform
illumination field is created using a special design diffusing
media, item 11 in FIG. 3. The diffusing media is fabricated from a
cube of optical grade polycarbonate or acrylic. As can be seen in
the Figure the test sample is placed along the centerline of the
illumination module. A cutout shaped like an elongated "U" is made
in one side and faces the optical sensor. The cutout is slightly
larger than the diameter of recessed bottom holder and test sample.
The LED illumination panels can be adjusted for backlight, diffuse
sidelight (forward scatter) or a combination of both. A voltage
controller allows the output of the LED lighting panels to set for
optimum contrast/performance. The front surface of the illumination
module, except for the viewing channel, is hidden by the aluminum
housing to protect over exposure of the sensor from the LED
lighting panels.
FIG. 4 illustrates the cut away view of the illumination module as
seen through section B-B of FIG. 3. The illumination diffuser and
LED panels extend below the bottom of the sample vessel to insure
uniform lighting across the entire image. This unique design hides
the corners of the LED panels and makes the entire illumination
field a uniform intensity. Variations in the height of sample
container are accommodated in the inspection position with the aid
of compliance spring (Item 13) and low friction guide (item 12).
Unlike previous designs this system allows particles in the
solution to be tracked throughout the entire volume. The technology
implements high resolution area scan sensors that acquire full
frame images in several milliseconds. The sensor is able to scan
the entire volume of the solution each frame. The detection of
particles>40 .mu.m are isolated with 100% certainty within the
inspection cycle.
The second major component in the inspection system is the Sensor
Module. The sensor module is designed as a sealed unit with no user
serviceable components. The image sensor, optics, filters are
pre-calibrated in known positions in the sensor module. FIG. 5
illustrates the front surface of the sensor module (item 16) and
the viewing window (item 17). The viewing window is constructed
using a material with anti-reflective coating. The window is sized
to accommodate the field of view (FOV) necessary to acquire the
image of the sample under inspection.
FIG. 6 illustrates one internal configuration of the Sensor Module.
The photosensitive detection system used in the sensor module is
either a high-resolution CCD sensor or in some applications a
sensitive CMOS sensor may be used. The CCD sensor (item 18) must be
of mega-pixel resolution or larger and is located in one corner of
the sensor module. The optical system is very important in the
detection of small particles in solution. High quality lenses
should be used to enhance performance of the inspection (item 19).
The optical path length (the distance between the CCD sensor and
the sample under inspection) has an effluence of the imaging
characteristic and performance of the system. In some cases the
path length must be longer than the available distance between the
physical location of the CCD sensor and the sample under
inspection. In this case, a folded optical path is employed by
reflecting an image of the object through one or more mirrors to
increase the apparent distance between object and CCD sensor as
illustrated by items 20 in FIG. 6. The longer the focal length of
the lens the greater the depth of field and therefore the larger
the volume that can be inspected. When instrument volume is at a
premium the folded optical path allows for better system
performance in a small footprint enclosure.
The sensor module incorporates an internal optical filter wheel.
The wheel is a disk with one or more filters (polarizing, grayscale
attenuation or color) that allow the system to change the CCD
sensor characteristics very rapidly. The filter wheel is
illustrated as item 21 in FIG. 6. The filter wheel is optional and
is not required for every inspection. The filter wheel is driven by
a small stepper or servomotor (item 22) from inside the sensor
module. The filter wheel may be substituted with a liquid crystal
window in grayscale applications and has the benefit of not having
a mechanical moving components. The liquid crystal window
attenuates the amount of light allowed to pass in the optical path.
This ability to attenuate the optical path, whether electronically
or mechanically, is critical in the inspection application. The
inspection process will be discussed later in this document.
The sensor module is a seal box with all optical devices mounted
inside. The sensor connections are made by way of a multi-pin
connector on the rear of the module. The multi-pin connector system
allows the user to easily replace a defective sensor module with
another sensor module that is pre-configured for the application
with no user setup required. When the initial application is
installed it defines the configuration of the sensor module. This
configuration is archived at the plant of manufacture so that an
exact duplicate sensor module can be assembled for use as required.
On the bottom of the sensor module is a pair of holes designed to
accept mating tapered dowel pins (item 25 of FIG. 8). The dowel
pins only allow the sensor module to be installed in a specific
location in the inspection system. The multi-pin connector is used
to connect the sensor and aperture control (liquid crystal window
or filter wheel) inside the sensor module without having the user
open the enclosure.
The relative position of components with respect to each other is
critical for system operation. The locations are defined by each
application. Enhancement have been made to the interior of the
sensor module to allow each unique configuration to be setup easily
and quickly. The CCD sensor is mounted on one or more dovetailed
slides that permit the unit to translate in orthogonal directions
as indicated by item 28 in FIG. 8. The dowel pins insure that the
sensor module is mounted the proper distance from the object under
inspection (item 29). The front a surface mirrors used to guide the
optical path use goniometer mountings for fine alignment of the
field of view to the target position (items 26 & 27). The
region of interest (ROI) when inspecting solution filled
pharmaceutical vials is from the bottom of the meniscus to the
bottom of the vial as illustrated by item 29. The solid angle of
the optical path defines the FOV of the image and is determined
primarily by the focal length of lens used, identified as item 30.
The solid angle of the optical path must be clear of
obstructions.
FIG. 9 illustrates the complete configuration of a single
inspection cell. The key components Illumination and Sensor Modules
are mounted on a flat tabletop or work piece. The working distance
and angle of viewing of the inspection cell is defined by the
distance between the axis of rotation of the rotational drive (item
33) and the dowel pins (item 32). The object height above the work
plane (item 31) is defined by the height of the recessed container
holder mounted on top of the drive shaft. A word should be said
about the rotational drive (item 33). The method of rotation is not
as important as the parameters used to perform the function. The
best results are achieved with a drive system that is capable of
accelerating and decelerating quickly. The physics of the
inspection require that the drive system accelerate rapidly,
maintain a constant velocity and then decelerate rapidly. The
profile of the motion curve is very important and defines the
motion or path of the contaminating particle in the solution. The
wall of the vessel must couple with the solution within. It is
important the acceleration/velocity profile does not cause
cavitation (the generation of air bubbles in the solution). If
cavitation is the result of the motion profile the sample can not
be inspected reliably. The motion profile must move the heavier
particles without allowing the meniscus to creep up the walls to
the vial neck. If the vial is spun too vigorously the particle may
be spun up into the cap of the container and be held there. The
correct motion profile of an inspection is defined by the
size/shape of the container and the viscosity of the solution
inside it. This invention allows the user to study the shape
characteristics of the meniscus while defining the motion
profile.
The Illumination Module is mounted on a linear translator that
allows it to be raised and lowered. Raising the Illumination Module
provides clear access to the sample container and rotational
drive/recessed holder. The linear translator (item 34) is normally
positioned at the rear of the Illumination Module. This has the
additional benefit of reducing the spacing between adjacent
inspection units if more than one is implemented. The linear
translator implementation can be assisted by air (cylinders),
electric (or magnetic), or mechanical (lead screws or cams). The
linear translator should be parallel to axis of rotation.
The last key component in the inspection system is the Image
Processor and Specialized Vision Software. The Sensor Module sends
image data (optical picture in electronic format) to the Image
Processor (item 35). The image processor acquires high resolution
(minimum 1280.times.1024 pixels) with a minimum signal to noise of
10. bits (1024 grayscale levels). Much higher resolution sensors
may be used when cost or cycle times at not as critical. The
preferred data transport mechanism is to use the Camera-Link (CL)
format indicated as item 37. The analysis of the image data is
performed using special software written to extract the particles
in solution. The system acquires multiple HR images in rapid
secession (4-60 images) and stores them in separate frame buffers.
The sensor acquisition control allows the application to define the
region of interest (ROI) from within the field of view (FOV). The
system should use frame rates (number of full pictures per second)
in the range of 24-60 frames per second. If partial frames are used
to acquire images with smaller field of view the frame rates
increase. The optimum frame rate is one in which the largest
diameter particle (assuming spherical object) translates or moves
at least one diameter between successive images. It may be the case
that the viscosity or fluid motion is slow and a delay must be
placed between successive image acquisitions. The software compares
each image with the previous image (except in the case of the
first) and isolates any object with the image field of view that
moves. A more advanced approach is for the software to compare each
image to a specified image in the acquisition sequence so that the
relative movement of the particle(s) can be very small and still be
detectable. This is important when detecting the motion and then
sizing of heavy or large particles that tend to settle very
rapidly. A special image-processing algorithm is used to extract
the moving particles and then determine their relative size.
The Image Processor (item 35) acts as the inspection cell master
controller and controls the other modules or devices in the
inspection cell. The Motor Controller (item 36) is used to generate
the motion profile in conjunction with the rotation drive. The
request to perform a motion profile is given to the motor
controller over item 38. The control line between the motor
controller and the motor is indicated by item 39. In the evaluation
unit a high torque stepper motor with lower inertia was used to
rotate the test sample. The motor controller also controls turning
the various LED lighting panels on or off during the inspection
(item 40). When the motion profile has been completed the motor
controller reports back to the image processor and the image
processor begins acquiring the necessary images. Depending on the
number and size of image acquired the entire inspection cycle
requires from one to several seconds.
If the average cycle time is three seconds for a rigorous
inspection then the inspection cell is limited to 20 inspections
per minute. The Laboratory Assay System is a small single
inspection cell unit designed to handle a limited number of samples
per hour. This does not lend itself toward mass production
inspection. However, the design concept can easily be expanded to
incorporate multiple copies of the inspection cell. FIG. 10
illustrate an approach that can handle the desired volume by
implementing multiple inspection cells side by side. The inspection
cell is indicated as item 45 and is comprised of a Sensor Module
(item 41), an Illumination Module (item 42), an Image Processing
Module (item 43) and the sample on rotational drive (item 44).
The large volumes of sample product would be moved into the
inspection position this may be performed at all stations
simultaneously if desired. However simultaneous operation is not
necessary as each inspection cell is independent. The simultaneous
operation would reduce the cost of the rotational motion by using a
common drive mechanism.
If a simultaneous operation were used the steps would include,
1. Loading of sample into the spin station
2. Select illumination configuration
3. Run samples through the motion profile
4. Acquire necessary images
5. Turn off illumination
6. Start samples unloaded while simultaneously start analysis of
images
7. Analyze images for possible particle defects and report
findings
8. Flag reject samples containers
9. Repeat sequence as required
It would be difficult to hand-load the laboratory assay system at
20 vials per minute. However, if 10 stations were used in a large
volume production system it would be easy to achieve 200 samples
per minute. The key feature of this inspection technology is the
ability to determine the size of particle inspection with an
accuracy range of 20 .mu.m when examining a 2-10 ml sample. The
user can select an exact cut-off limit below which particles
smaller than the limit will be accepted. The product is not
rejecting simply on a detection basis but on a particle size
basis.
When calibrated using NIST traceable standard samples the
inspection system provides a method for validation for maximum
dimensional particle sizing. This also provides a more realistic
measurement of non-spherical particles like platelets, fibers and
non-uniform shapes (glass shards). The Module concept provides NIST
traceable inspection not only when shipped but virtually forever.
This is possible because of a stable detector with permanent size
calibration.
* * * * *